236 Proceedings of the Royal Society of Edinburgh, [Sess. 
The resistance was then calculated from the formula 
1047 + d 
1047 - d 
x S 
where S = 394*224(1 + 0*00015 (t 15°)} + 0*006 (the standard coil and its leads), 
and R = the resistance of the water and its lead. Hence the actual resistance 
of the water is R — 0*0669 legal ohms, the reciprocal of which multiplied 
by the constant of the cell ( vide supra) gives the true specific conductivity. 
Errors of Experiment . — So far as the author is aware, the only errors 
not common to all the ninety measurements are those incidental to correct 
balancing and reading of the bridge. And the extreme limit would be half 
a millimetre of the wire. As the mean of several readings was taken, it 
would probably be much less than this ; but allowing 0*25 mm. of wire as a 
possible error, we find that this affects the conductivity by 2 in the fifth 
figure, an amount which corresponds to the effect produced by an error of 
0*01 grams in the estimation of chlorine. 
The Optical Measurements. 
These were made by means of a refractometer reading directly to 
30" and which could be easily estimated to 15". The prism was a glass 
one with movable sides made by Hilger, and when set up had an angle 
of 60° 6' 10". The prism was not broken till No. 87, when it was reset and 
the angle found to be 60° 3' 54". The values for the last four measure- 
ments were calculated for a prism of 60° 6' 10", and are so given. 
Methods of Experiment . — The bottles containing the water were kept 
near the refractometer, and were thus at the temperature of the room. 
The actual temperature of the water in the prism was determined by a 
thermometer reading to tenths of a degree, which had been standardised by 
comparison with a Kew standard. The left-hand edge of the image of the 
slit was made to coincide with the point of intersection of the cross wires. 
Sodium light was used as illuminant. The prism was rinsed twice with 
the water to be examined, and the thermometer inserted. The prism was 
placed in position on the refractometer and the position (1) of minimum 
deviation found and read ; the temperature of the water was then read, 
the prism was turned round through 180°, and the position (2) of minimum 
deviation again read. The whole operation was then repeated. 
Half the difference between 1 and 2 gives the angle of minimum devia- 
tion at the temperature noted. Corrections for temperature were applied 
as below, and the mean of the two values taken as correct. The difference 
in no case exceeded 30". In two cases only it exceeded 15". 
